Paper Reading
文章平均质量分 52
zenHorse
Deep Learning Computer Vision
展开
-
Feature Selective Networks for Object Detection
Main Contributions:sub-region attention map and aspect ratio attention map for each RoI RoI feature extractoradopt a 1 × 1 convo- lutional layer to reduce the channel number to Cs and pool...原创 2018-08-28 09:45:15 · 947 阅读 · 0 评论 -
Single-Shot Object Detection with Enriched Semantics
Detection with Enriched Semantics (DES) is a singleshot object detection network with three partsa single shot detection branch, ssd a segmentation branch to enrich semantics at low level dete...原创 2018-08-28 09:49:06 · 757 阅读 · 0 评论 -
Repulsion Loss Detecting Pedestrians in a Crowd
Contritutions:study the impact of crowd occlusion on pedestrian detection Two types of repulsion losses: RepGT Loss and RepBox Loss a crowd-robust pedestrian detector is trained end-to-endLo...原创 2018-09-04 15:22:25 · 609 阅读 · 0 评论 -
UnitBox An Advanced Object Detection Network
where xt, xb, xl , xr represent the distances between curerent pixel location (i, j) and the top, bottom, left and right bounds of ground truth, respectivelyL2 Loss Layer:two drawbacks...原创 2018-09-04 18:35:03 · 884 阅读 · 0 评论